Chemosphere 161 (2016) 349e357

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Modeling the oxidation of phenolic compounds by hydrogen peroxide

photolysis

ez*
Tianqi Zhang, Long Cheng, Lin Ma, Fanchao Meng, Robert G. Arnold, A. Eduardo Sa
Department of Chemical and Environmental Engineering, University of Arizona, Tucson, AZ, 85721, United States

h i g h l i g h t s

A mechanism for oxidation of phenolic compounds by hydroxyl radicals is presented.

Hydroxyl radical scavenging by intermediates reduces effectiveness of process.

Kinetic model of the full mechanism is necessary for prediction of reactor performance.

a r t i c l e i n f o a b s t r a c t

Article history: Hydrogen peroxide UV photolysis is among the most widely used advanced oxidation processes (AOPs)
Received 7 March 2016 for the destruction of trace organics in waters destined for reuse. Previous kinetic models of hydrogen
Received in revised form peroxide photolysis focus on the dynamics of hydroxyl radical production and consumption, as well as
13 June 2016
the reaction of the target organic with hydroxyl radicals. However, the rate of target destruction may also
Accepted 29 June 2016
be affected by radical scavenging by reaction products. In this work, we build a predictive kinetic model
for the destruction of p-cresol by hydrogen peroxide photolysis based on a complete reaction mechanism
Handling Editor: I. Cousins that includes reactions of intermediates with hydroxyl radicals. The results show that development of a
predictive kinetic model to evaluate process performance requires consideration of the complete reac-
Keywords: tion mechanism, including reactions of intermediates with hydroxyl radicals.
Advanced oxidation processes 2016 Elsevier Ltd. All rights reserved.
Kinetic model
Phenolic compounds
Radical reactions

1. Introduction
Alkylphenol ethoxylates are non-ionic surfactants commonly
used to formulate products such as detergents, paints, dispersing
agents, wetting products, pesticides, and lubricants (Nagarnaik and
Boulanger, 2011). These compounds are converted into alkylphe-
nols in bioreactors used in wastewater treatment (Ahel et al., 1996),
leading to their presence in surface waters (Scullion et al., 1996;
Dong et al., 2015), groundwater (Swartz et al., 2006), municipal
and industrial wastewater (Loyo-Rosales et al., 2007), aquatic
sediments and marine shellsh (Lye et al., 1999). The most
commonly detected metabolite, p-nonylphenol, induces breast tu-
mor cell proliferation (Soto et al., 1991) and is a recognized endo-
crine disrupter (Lee and Lee, 1996). Therefore, the fates of p-
nonylphenol and other phenolic compounds during wastewater
* Corresponding author.

ez).
E-mail address: esaez@email.arizona.edu (A.E. Sa
treatment and following the release of treated wastewater into the
environment is of practical importance.
Advanced oxidation processes (AOPs) are effective treatments
for phenolic compounds due to their high rates of reaction with
hydroxyl radicals (second-order rate constants in the range
109e1010 M1 s1; Buxton et al., 1988). The UV photolysis of
hydrogen peroxide (UV/H2O2) is among the most widely used AOPs
in water treatment. Various models have been developed to
simulate process kinetics. Common elements in these models
include: (i) the kinetics of radical production from the direct
photolysis of hydrogen peroxide, and hydroxyl radical scavenging
by hydrogen peroxide and other radicals; (ii) reaction of hydroxyl
radicals with the target compound and, although to a lesser extent,
(iii) radical scavenging by reaction byproducts and intermediates
(Lay, 1989; Glaze et al., 1995; Liao and Gurol, 1995; Stefan et al.,
1996; Crittenden et al., 1999; Murcia et al., 2015; Rojas et al.,
2010, 2011).
The rst step in the reaction mechanism for radical oxidation of
phenolic compounds is OH addition to the ring (Wojn
arovits et al.,

http://dx.doi.org/10.1016/j.chemosphere.2016.06.110
0045-6535/ 2016 Elsevier Ltd. All rights reserved.
350 T. Zhang et al. / Chemosphere 161 (2016) 349e357
2002). The position of OH addition is controlled by substitutions on
the ring, but para- and ortho-hydroxylation normally predominate
due to the electrophilic nature of the hydroxyl radical (Omura and
Matsuura, 1968; Matsuura and Omura, 1974). Further oxidation of
the resulting dihydroxy compound occurs via abstraction of a
hydrogen atom to form benzoquinones, after which the ring can be
cleaved to yield aliphatic acids (Scheck and Frimmel, 1995).
Previous efforts on the kinetic modeling of the UV/H2O2 process
generally include a complete description of the chemistry behind
radical formation, but rarely consider reaction intermediate and by-
products. Glaze et al. (1995) proposed a kinetic model containing
the most important elements of UV/H2O2 photolysis to predict the
time-dependent concentration of a halogenated target. They used a
quasi-steady-state assumption (QSSA) to calculate the time-
dependent concentrations of free radicals. Their model did not
account for CO2 evolution/pH change and neglected the role of all
radical scavenging species except H2O2. Stefan et al. (1996) fol-
lowed a similar approach, but assumed that the target compound
(in their case acetone) was transformed to oxalic and formic acids
and then to CO2. Subsequently, Crittenden et al. (1999) modeled
UV/H2O2 process without reliance on a quasi-steady-state
assumption for radical concentrations. Their work showed that
the QSSA tends to under-predict hydroxyl radical concentrations in
some cases and therefore the rates of target compound decompo-
sition. Furthermore, their model included radical scavenging
reactions.
The model of Song et al. (2008) accounted for time-dependent
radical concentrations, pH changes due to product formation (CO2
and low molecular weight acids), and radical scavenging by H2O2
and natural organic matter.
More recent modeling efforts of oxidation of phenolic com-
pounds usually include a full formulation of the dynamics of radical
formation, but very few or no details of the inuence of interme-
diate compounds and by-products over target compound degra-
dation. De Luis et al. (2011) presented a model of phenol and cresol
degradation during the UV/H2O2 process that considered possible
reactions of the organic compounds with produced radicals, but
without considering intermediate compounds and by-products. A
similar approach was followed by Rojas et al. (2010, 2011) to model
the UV/H2O2 and Fentons reaction processes. Moreira et al. (2012)
developed a phenomenological model to simulate the TiO2 pho-
tocatalysis of phenol, which includes specic details about the re-
action mechanism.
In this work, p-cresol was selected as an example of a
substituted phenolic compound. Even though p-cresol itself is not a
common trace organic in wastewater efuent, it has chemical
structure and reactivity with hydroxyl radicals similar to p-non-
ylphenol, which is a common endocrine disrupter compound
widely present in wastewater efuent. A mechanism for radical-
dependent UV/photolysis was proposed and calibrated based on
time-dependent measurements of p-cresol and a number of reac-
tion products/intermediates in a batch reactor. The model accounts
for changes in ionic strength and pH due to reactant conversions.
Reactor uence rate and the p-cresol quantum yield during direct
photolysis were considered in the kinetic model. The calibrated
model was employed to predict the evolution of concentrations of
p-cresol and reaction intermediates under various experimental
conditions.
2. Materials and methods
2.1. Chemicals
All chemicals were obtained from commercial sources and used
without further purication, including p-cresol (Acros, 99%), 4-
methylcatechol (Acros, 98%), hydrogen peroxide (Acros, 50 wt%),
titanium sulfate (Pfaltz&Bauer, 30%), formate (Fluka, 1000 mg/
L 4 mg/L), acetate (Fluka, 1000 mg/L 4 mg/L), oxalate (Fluka,
1000 mg/L 4 mg/L), and TOC standard (Fluka, 50 mg/L 2 mg/L).
Methanol (Fisher Chemical, HPLC Grade) and acetonitrile (Fisher
Chemical, HPLC Grade) were solvents for HPLC analyses. Milli-Q
water (resistivity 18.0 mU cm1) was obtained from a Barnstead
NANOpure II system. Glassware was washed and heated overnight
at 550  C prior to use.
2.2. Experimental procedure
All experiments were conducted in a cylindrical (14-cm diam-
eter, 950-mL) glass batch reactor, equipped with low pressure (LP)
UV lamp, UV uence meter (Hamamatsu, H8025-254) and mag-
netic stirrer. Stock solutions of p-cresol were prepared one day
before use. 4-Methylcatechol stock solutions were prepared a half
hour before use in amber glass containers. The UV lamp (43-cm
long) was turned on two hours before the experiments. The lamp
position was 21 cm above the liquid level in the reactor. A quartz
sleeve covered the lamp to cut-off wavelengths below 185 nm.
Solution volume was 300 mL in all experiments (liquid depth was
1.95 cm). Experiments were initiated by exposing the reactor to UV
light. Experiment duration depended on the initial p-cresol con-
centration but ranged from one to three hours. Samples were
analyzed immediately after withdrawal. Dilution was frequently
necessary to maintain concentrations within the recommended
analytical range. Temperature was maintained at T 24.5 0.5  C.
The trajectory of solution pH was determined in each experiment.
2.3. Analytics
The concentration of p-cresol was measured by uorescence
spectrometry (PerkinElmer LS 55). The inner lter effect (IFE)
caused by light absorption and reabsorption of light-absorbing
compounds was corrected mathematically, accounting for the ab-
sorption spectrum of the sample (MacDonald et al., 1997). Solution
spectra were measured using a UVeVis spectrophotometer
(Thermo Science, Genesys 10s). Even though the reaction inter-
mediate 4-methylcatechol is produced in concentrations that are
comparable in order of magnitude to p-cresol, independent mea-
surements showed that its uorescence intensities were negligible
compared to those of p-cresol in the relevant range of wavelengths.
Analysis of primary intermediates, particularly 4-methyl-
catechol, was carried out by HPLC (Agilent Technologies, 1200 Se-
ries) with a Polar-RP, 80 A column (15 cm  4.6 mm  4 mm)
(Phenomenex) and UV/VIS detector at 280 nm. Elution solvents
were 10% methanol in water (A) and acetonitrile (B), Q 0.4 mL/
min. The mobile phase initially consisted of 45% vol. B, linearly
ramped to 65% vol. B at 15 min and then restored to the initial
condition.
Formate, acetate and oxalate were measured via ion chroma-
tography (Dionex, ICS-5000) equipped with an AS18 column
(2  250 mm) (Dionex IonPac), an AG18 guard column
(2  50 mm) (Dionex IonPac), and an AS-DV (Dionex) autosam-
pler. Ions were detected by suppressed conductivity of the eluent
using an ASRS-2mm self-regenerating suppressor (Dionex). The
eluent (KOH solution; Q 0.4 mL/min) was programmed as fol-
lows: 1 mM for 8 min, ramped (linearly) to 30 mM in 20 min, then
immediately restored to 1 mM.
Hydrogen peroxide was measured using a modied perox-
ytitanic (colorimetric) method (Boltz and Holwell, 1978). Absor-
bance was measured spectrophotometrically (Thermo Science,
Genesys 10s) at 407 nm. Samples were diluted to maintain absor-
bance values within linear limits. TOC was monitored using a Total
 T. Zhang et al. / Chemosphere 161 (2016) 349e357
Organic Carbon analyzer (Shimadzu, TOC-VCSH).
2.4. Proposed mechanism
The proposed mechanism of p-cresol oxidation by hydroxyl
radicals is shown in Fig. 1. The rst step is OH radical addition to the
ring to form 4-methylcatechol (Matsuura and Omura, 1974; Olariu
et al., 2002). Two possible pathways (Omura and Matsuura, 1968;
Scheck and Frimmel, 1995) are used to explain the hydroxylation.
One is the attack of OH radical on the phenolic ring, resulting in the
formation of a cyclohexadienyl radical, which is subsequently
converted into 4-methylcatechol. The other involves abstraction of
a hydrogen atom yielding a phenoxyl radical that can combine with
a second OH radical. However, the hydrogen abstraction reaction of
a methyl H atom (Wang et al., 1998) is at least one order of
magnitude slower than the formation of a cyclohexadienyl radical
(Sehested et al., 1975; Wojna
rovits et al., 2002). Further hydroxyl-
ation may convert 4-methylcatechol; however, this reaction has
negligible yield (Scheck and Frimmel, 1995). Oxidation of 4-
methylcatechol to 4-methylbenzoquinone occurs in two steps:
the OH radical attacks the ring forming a 4-methylcatechol semi-
quinone radical (Gohn and Getoff, 1977), after which the semi-
quinone radical converts to 4-methylbenzoquinone in the presence
of molecular oxygen (Davies et al., 1995). Once 4-methyl-
benzoquinone is formed, ring tension and oxidizing agents force
ssion of the ring. The CeC bond in the ring between the carbons
attached to oxygen atoms is cleaved, and 3-methyl-muconic acid is
one of the ssion products (Pospisil et al., 1957; Scheck and
Frimmel, 1995). Oxidation of 3-methyl-muconic acid involves the
carbon double bonds. Because of the higher acidity of the a-
hydrogen adjacent to the double bond, it can be easily abstracted by
the OH radical (Jin et al., 2010; Huang et al., 2011). Low molecular
weight organic acids such as lactic and oxalic acid are oxidation
products. The abstraction of a-hydrogen next to a eCOOH group by
OH radical forms additional low-molecular -weight acids such as
acetic and formic acids (Boonrattanakij et al., 2009). Eventually,
oxalic acid, acetic acid and formic acid are oxidized by OH radicals
Fig. 1. Proposed mechanism of p-cresol oxidation by hydroxyl radicals (see text for description).
351
to carbon dioxide (Getoff et al., 1971; Leitner and Dore, 1997; Hart,
1951; Sun and Saeys, 2008).
2.5. Kinetic model
The proposed kinetic model for the oxidation of p-cresol during
UV/H2O2 photolysis consists of 36 reactions (Table 1). The UV/H2O2
photolysis mechanism, including an initiation step that produces
hydroxyl radical (OH) and propagation steps that scavenge OH
and produce other active radicals (O 
2 and CO3 ), is based on the
work of Rojas et al. (2010, 2011) and references therein. Since the
reactor was open to the atmosphere, the initial carbon dioxide
concentration was assumed to be at equilibrium with atmospheric
carbon dioxide. The scavenging effect of inorganic carbon species,
bicarbonate and carbonate ions was considered in the model, but
calculations showed that their effect was negligible in the range of
conditions of the experiments.
The rate of direct photolysis of H2O2 is given by
RR1 FH2 O2 fH2 O2 I0 e2:303Az (1)
where FH2 O2 (Table 1) is the quantum yield for the photocatalytic
decomposition of H2O2, fH2 O2 is the fraction of total light absorbed
by H2O2, I0 is the average uence rate (in Einsteins per unit time
and reactor volume) at the reactor surface and Az is the total
absorbance of the solution as function of depth. In the model, H2O2,
PC, 4 MC and 4 MB are the only important light-absorbing species
at 254 nm. Therefore, Az is given by


Az H2 O2 H2 O2  PC PC 4MC 4MC 4MB 4MB z
(2)
where H2 O2 ; PC ; 4MC (Table 1) and 4MB (2100 M1 cm1; Waite,
1976; Jongberg et al., 2011) are molar extinction coefcients for
hydrogen peroxide, p-cresol, 4-methylcatechol and 4-methyl-
benzoquinone at 254 nm, respectively, and z is the distance
measured downward within the reactive mixture from the top
352 T. Zhang et al. / Chemosphere 161 (2016) 349e357

Table 1
Kinetic and equilibrium parameters used in the model. Acronyms are dened in Fig. 1. Quantum yield (F) and molar extinction coefcients () are reported for photolysis
reactions, kinetic rate constants (k) are reported for relevant reactions, and equilibrium constants (K) are reported for reactions assumed to reach equilibrium instantaneously
in the reaction mixture.

Reaction Parameters Reference

1 1
R1 H2O2 hn / 2 OH 1 18.7 M cm Volman and Chen (1959)
F1 0.5 mol/Ein

R2 OH H2O2 / O 2 H2O H

k2 2.7  107 M1 s1 Buxton et al. (1988)

R3 OH HO 
2 / O2 H2O k3 7.5  109 M1 s1 Buxton et al. (1988)

R4 OH HCO 
3 / CO3 H2O k4 8.5  106 M1 s1 Buxton et al. (1988)
 
R5 OH CO2
3 / CO 3 OH k5 3.9  108 M1 s1 Buxton et al. (1988)

R6 OH HO2 / O2 H2O k6 6.6  109 M1 s1 Buxton et al. (1988)

R7 OH O
2 / O2 OH

k7 8.0  109 M1 s1 Buxton et al. (1988)

R8 OH OH / H2O2 k8 5.5  109 M1 s1 Buxton et al. (1988)

R9 OH CO
3 / products k9 3.0  109 M1 s1 Holeman et al. (1987)
R10 O
2 H2O2 / OH O2 OH

k10 0.13 M1 s1 Bielski et al. (1985)
 
R11 O2 CO3 / O2 CO2 3 k11 6.5  108 M1 s1 Eriksen et al. (1985)
R12 O 
2 HO2 / HO2 O2 k12 9.7  107 M1 s1 Bielski et al. (1985)
R13 HO2 HO2 / H2O2 O2 k13 8.6  105 M1 s1 Weinstein and Bielski (1979)
R14 HO2 H2O2 / OH O2 H2O k14 3.7 M1 s1 Bielski et al. (1985)
R15 CO 
3 H2O2 / HCO3 O2 H

k15 8.0  105 M1 s1 Neta et al. (1988)
R16 CO 
3 HO2 / HCO3 O2
 
k16 3.0  107 M1 s1 Neta et al. (1988)
R17 CO 
3 CO3 / 2CO3
2
k17 2.0  107 M1 s1 Neta et al. (1988)
R18 H2O2 4 HO 2 H

k18 2.51  102 s1 Perry et al. (1981)
k18r 1  1010 M1 s1
R19 HO2 4 O
2 H

k19 1.58  105 s1 Perry et al. (1981)
k19r 1  1010 M1 s1
R20 CO2(aq) H2O 4 H2CO3 k20 0.039 s1 Housecroft and Sharpe (2005)
k20r 23 s1
R21 H2CO3 4 HCO
3 H
k21 1.25  107 s1 Greenwood and Earnshaw (1997)
k21r 5  1010 M1 s1
R22 PC hn / products 22 340 M1 cm1 This work
F22 0.034 mol/Ein
R23 PC OH / 4 MC k23 1.2  1010 M1 s1 Buxton et al. (1988)
R24 4 MC hn / products 24 310 M1 cm1 This work
F24 0.007 mol/Ein
R25 4 MC OH / 4 MB k25 1.6  1010 M1 s1 Gohn and Getoff (1977)
R26 4 MB OH / 3 MM k26 2.0  1010 M1 s1 This work
R27 3 MM OH / L 2 Ox (Ox2) k27 5.0  108 M1 s1 Kang et al. (2002)
R28 L OH / HA (A) CO2 H2O k28 3.0  108 M1 s1 Buxton et al. (1988)
R29 Ox OH / Ox H2O k29 4.7  107 M1 s1 Buxton et al. (1988)
R30 Ox2 OH / Ox OH k30 7.7  106 M1 s1 Buxton et al. (1988)
R31 Ox Ox / Ox (Ox2) 2 CO2 k31 5.0  108 M1 s1 Ershov et al. (2008)
R32 HA OH / F CO2 H2O k32 1.23  108 M1 s1 Sun and Saeys (2008)
R33 A OH / F CO2 OH k33 8.5  107 M1 s1 Buxton et al. (1988)
R34 F OH / CO2 H2O k34 9.77  107 M1 s1 Sun and Saeys (2008)
E35 Ox 4 Ox2 H K35 6.31  105 M Benjamin (2002)
E36 HA 4 A H K36 1.74  105 M Benjamin (2002)
E37 HCO 2
3 4 CO3 H

K37 4.48  1011 M Benjamin (2002)

surface. Note that Acid-base reactions are considered to be at equilibrium. The

equilibrium equations are used to relate concentrations of the
H2 O2 zH2 O2  species involved. For example, for acetic acid,
fH2 O2 (3)
Az
The volume averaging of equation (1) over the batch reactor is,
after manipulations,  
TOTA g1 H
  HA   (6)
RR1 FH2 O2 fH2 O2 I0 1  e2:303A (4) K36 g1 H

where the angular brackets denote volume average, and

A Az z l, where l is the depth of the reacting mixture. Other
direct photolysis reactions are treated in the same manner.
TOTA K36
Radical reactions are generally second-order reactions. For g1 A    (7)
K36 g1 H
example, the reaction rate for the consumption of p-cresol by OH
radical is
where TOTA HA A , K36 is equilibrium constant of acetic acid
h i dissociation and g1 is the activity coefcient, which is calculated
_
RR23 k23 PC OH (5) from the Davies Equation.
Based on the mechanism proposed, the corresponding mole
where k23 is the second-order rate constant of p-cresol reaction balances of H2O2, p-cresol and 4-methylcatechol yield the following
with OH (Table 1). system of differential equations:
 T. Zhang et al. / Chemosphere 161 (2016) 349e357 353

dH2 O2    h i
FH2 O2 fH2 O2 I0 1  e2:303A  k2 H2 O2  OH _
dt
h i2 h i h i2
k8 OH_  k10 H2 O2 g1 O_ _
2 k13 HO2
h i h i
 k14 H2 O2  HO_2  k15 H2 O2 g1 CO 3
_

  h i
 k18 H2 O2  k18r g1 HO 2 g1 H

(8)

dPC   h i
_
FPC fPC I0 1  e2:303A  k23 PC OH (9)
dt

d4MC   h i
_
F4MC f4MC I0 1  e2:303A k23 PC OH
dt
h i
 k 4MC OH _ (10)
25

TOC TOC0  CO2  CO2 0 (11)

Fig. 2. Decomposition of hydrogen peroxide in Milli-Q water in the batch reactor. Solid
where TOC0 and CO2 0 are initial TOC and CO2 molar concentra- lines: model predictions obtained by tting the model to decay curves using the u-
tions. More details of the model formulation, including mole bal- ence rate as adjustable parameter.
ances of radicals (OH, HO2, O 2 and CO
3 ) and reaction
intermediates (4 MB, 3 MM, LA , OA , OA2, OA, AA, AA and
 

FA), plus algebraic equations ([H2CO3] as a function of [TOTCO3]
and charge balance) are provided in the supporting information.
The fact that the reactor is well mixed implies that most species
concentrations are uniform in the reactor. However, the relatively
high reaction rates of radical species suggest that their concentra-
tions are not necessarily uniform. For example, the OH radical
production by reaction R1 (Table 1) is not uniform due to the
attenuation of light with depth. In the model equations, radical
concentrations are volume-averaged concentrations. The formula-
tion is based on the assumption that kinetics that are not linear on
radical concentrations (such as in radical/radical reactions like R8,
Table 1) can be expressed in terms of volume-averaged radical
concentrations. This approximation was explored in detail by Rojas
et al. (2010).
The nal system of equations was solved using the stiff-
differential-equations solver ODE15s in MATLAB (based on Gears
method). The relative tolerance was 108, whereas the absolute
tolerances for concentrations of all species were set four orders of
magnitude below their respective values.
3. Results and discussion
3.1. Actinometry. Hydrogen peroxide photolysis
The average uence rate at the surface of the reactor was
determined using H2O2 actinometry. Hydrogen peroxide direct
decomposition experiments were conducted at various initial
concentrations. The kinetic model was applied to calculate the
reactor surface average uence rate by tting experimental data for
H2O2 decay. Reactions 1e21 (Table 1) were used in the model
formulation. A value of I0 0:85  106 Ein L1 s1 was obtained
(Fig. 2), which was in agreement with the uence rate measured by
the uence meter.
3.2. p-Cresol and 4-Methylcatechol direct photolysis
Quantum yields of p-cresol and 4-methylcatechol were deter-
mined from direct photolysis experiments for each compound
separately. In the experiments, p-cresol or 4-methylcatechol solu-
tions were exposed to UV light in the absence of H2O2. The quan-
tum yield of each compound was calculated by tting the kinetic
model to data representing compound decay. The solution absor-
bance for all experiments was measured as a function of time to
determine Az (equation (2)). This was necessary in direct photolysis
experiments since data showed that the products of direct
photolysis reaction also absorb 254-nm light. (2,20 -Dihydroxy-5,5-
dimethylbiphenyl, 4-methylcatechol and 2-hydroxy-40 ,5-dime-
thyldiphenyl ether have been reported as products in the direct
photolysis of p-cresol, Joschek and Miller, 1966.) Results are pre-
sented in Fig. 3. To generate model predictions, we used the
calculated uence from the actinometry experiments (Fig. 2) and
tted the quantum yields to experimental data.
3.3. Model predictions: p-cresol degradation
Fig. 4 provides predicted and experimental decay curves for p-
cresol in a series of experiments carried out at different initial
concentrations of H2O2 and p-cresol. Model calculations include the
uence rate determined via actinometry and the quantum yields
for p-cresol and 4-methylcatechol determined via direct photolysis.
In all cases experimental results are accurately predicted by the
model. Omitting OH radical scavenging by reaction intermediates
signicantly overestimates the destruction rate of p-cresol in most
cases (dashed lines in Fig. 4), which highlights the importance of
incorporating the full mechanism in the kinetic model.
Calculations show that intermediates, especially those before
ring cleavage (4-methylcatechol and 4-methylbezoquinone) played
an important role in hydroxyl radical quenching. There is no re-
ported kinetic rate constant for the reaction of 4-methyl-
benzoquinone reaction with hydroxyl radical although Scheck and
Frimmel (1995) reported that 1,2-benzoquinone was not detected
during oxidation of phenol by the UV/H2O2 process because it is
easily cleaved into aliphatic compounds. In this study, the rate
constant for the reaction of 4-methylbenzoquinone with hydroxyl
radical was held to be of the same magnitude as that of 4-meth-
ylcatechol, k26 2:0  1010 M 1 s1 , near the diffusion limit. Model
results were obtained with no additional tted parameters; all
other rate constants were obtained from preliminary experiments
(i.e. actinometry and direct photolysis experiments) or published
data (Table 1).
Molecular oxygen participates in several steps of the reaction
mechanismdspecically, in the oxidation of 4-methylcatechol to 4-
354 T. Zhang et al. / Chemosphere 161 (2016) 349e357

Fig. 3. Decomposition of (a) p-cresol, and (b) 4-mehtylcatechol in aqueous solution by (LP)UV direct photolysis. Solid lines: model predictions obtained from a t of model results to
decay curves using the respective quantum yield as adjustable parameter. (a) [PC]0 0.438 mM, (b) [4 MC]0 0.103 mM.

Fig. 4. Degradation of p-cresol in aqueous solution by UV/H2O2 advanced oxidation as a function of initial H2O2 concentrations (a), and initial p-cresol concentrations (b). Solid
lines: full model predictions using the proposed mechanism; dashed lines: model calculations neglecting reactions of intermediates. (a) [PC]0 0.1 mM, (b) [H2O2]0 1 mM.

methylbenzoquinone (Davies et al., 1995) and the oxidation of
carbon double bonds (Jin et al., 2010; Huang et al., 2011). To test the
potential role of oxygen, we performed experiments with degassed
solutions by bubbling argon into the solution before exposure to UV
light. During the experiment, an argon atmosphere was maintained
over the reactor to prevent oxygen diffusion into the reacting
mixture. The results in terms of p-cresol concentration vs. time
showed no appreciable difference relative to the original experi-
ments, suggesting that dissolved oxygen is not important in p-
cresol oxidation. Representative results are provided in the sup-
porting information.
There are two possible pathways for the OH-induced oxidation
of p-cresol to 4-methylcatechol (Omura and Matsuura, 1968;
Scheck and Frimmel, 1995): the rst involves the cyclohexadienyl
radical formed due to OH radical attack that converts the p-cresol
into dihydroxy compounds. In the second, the phenoxyl radical
formed due to the abstraction of hydrogen atom subsequently
combines with a second OH radical. However, the second mecha-
nism is dominant only when the OH-radical is present at high
concentrations (Waters, 1964). In the kinetic model used here, only
one OH radical is involved in the conversion of p-cresol to 4-
methylcatechol, suggesting that the cyclohexadienyl pathway
predominates.
The production of 4-methylcatechol was measured in the
experiments (Fig. 5). The reaction of p-cresol with OH radical was
the only 4-methylcatechol source, but two sinks are considered in
the kinetic model: direct photolysis and reaction with OH radical.
Even though the rate constant for 4-methylcatechol reaction with
OH radical is higher than that of p-cresol, 4-methylcatechol accu-
mulates until most of the p-cresol is consumed. There was good
agreement between data and predictions at low H2O2 concentra-
tions (Fig. 5a), but the model tends to underpredict concentrations
of 4-methylcatechol at later times and higher measured concen-
trations (Fig. 5b) while still trending correctly.
The evolutions of formic acid, acetic acid and oxalic acid con-
centrations are presented in Fig. 6 for a representative experiment.
(Results for p-cresol degradation in this experiment are in Fig. 5b.)
Although the a-hydrogens next to the carbon double bonds and
carboxyl groups are more active under oxidizing conditions, there
may be several ways in which to break down the aliphatic acid
formed by the ring cleavage. Therefore, the nal stages of the
mechanism proposed (Fig. 1) may lead to formation of other spe-
cies. Fig. 7 shows the trajectory of total carbon corresponding to
measured concentrations of aliphatic acids along with model cal-
culations for a representative experiment. Both model and data
indicate that aliphatic acids accumulate throughout the experi-
ment. By the end of the experiment (180 min), essentially all the p-
cresol is destroyed (Fig. 5b).
 T. Zhang et al. / Chemosphere 161 (2016) 349e357
Fig. 5. Degradation of p-cresol in aqueous solution by UV/H2O2 advanced oxidation. Evolution of 4-methylcatechol concentration with time. Solid lines: model predictions.
[PC]0 0.5 mM, (a) [H2O2]0 1 mM, (b) [H2O2]0 5 mM.
Fig. 6. Evolution of formic acid, acetic acid and oxalic acid concentrations with time.
[PC]0 0.5 mM, [H2O2]0 5 mM.
Fig. 7. Evolution of aliphatic reaction intermediates concentration with time. Solid
line: model prediction. [PC]0 0.5 mM, [H2O2]0 5 mM. Abbreviations in Fig. 1.
355
The concentrations of H2O2 and TOC, and pH were measured
throughout the experiment and predicted using the kinetic model
(Figs. 8e10). Despite some differences, the model adequately pre-
dicts trends in H2O2 concentration (Fig. 8) and pH (Fig. 9). Previous
studies have calculated pH changes during this type of experiment
by assuming that the target organics are completely converted into
carbon dioxide (Crittenden et al., 1999; Rojas et al., 2010). In our
study, a signicant drop in pH was observed in the 3-h experiment.
However, the drop of pH was due to the formation of organic acids.
The lack of CO2 production (that is, complete mineralization of the
target) is evident in the TOC results (Fig. 10) since TOC does not
change appreciably during the experiments, as predicted by our
model.
3.4. Model predictions: 4-Methylcatechol degradation
Experiments were carried out with 4-methylcatechol as the
initial substrate at different initial hydrogen peroxide concentra-
tions. As anticipated, degradation rate was directly related to the
initial hydrogen peroxide concentration within the range of ex-
periments (Fig. 11). Agreement between experimental data and
model simulations is observed at 0.2 mM H2O2 but the model tends
Fig. 8. Evolution of H2O2 concentration with time. Solid line: model prediction.
[PC]0 0.5 mM, [H2O2]0 5 mM.
356 T. Zhang et al. / Chemosphere 161 (2016) 349e357
Fig. 9. Evolution of pH with time. Solid line: model prediction. [PC]0 0.5 mM,
[H2O2]0 5 mM.
Fig. 10. Evolution of TOC with time. Solid line: model prediction. [PC]0 0.5 mM,
[H2O2]0 5 mM.
to over-estimate degradation rates at higher H2O2 concentrations.
4. Concluding remarks
A comprehensive kinetic model was developed to quantify the
oxidation of p-cresol by UV/H2O2 AOP in a batch reactor. The model
was based on known elementary chemical and photo-chemical
reactions of the UV/H2O2 photolysis system, together with experi-
mentally established direct photolysis reactions and proposed OH-
radical reactions with intermediates. The kinetics of p-cresol
decomposition, as well as the main reaction intermediate (4-
methylcatechol) was successfully simulated throughout a wide
range of operating conditions. The approach employed shows that
the full reaction mechanism needs to be considered to evaluate
process performance. In particular, reaction intermediates, espe-
cially those formed before ring cleavage, play an important role in
the quenching of hydroxyl radicals and, consequently, those com-
pounds must be considered in kinetic models of AOP degradation of
phenolic compounds.
Fig. 11. Degradation of 4-methylcatechol in aqueous solution by UV/H2O2 advanced
oxidation as a function of initial H2O2 concentrations. Solid lines: model predictions.
[4 MC]0 0.1 mM.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.chemosphere.2016.06.110.
References
Ahel, M., Schaffner, C., Giger, W., 1996. Behaviour of alkylphenol polyethoxylate
surfactants in the aquatic environment - III. Occurrence and elimination of their
persistent metabolites during inltration of river water to groundwater. Water
Res. 30, 37e46.
Benjamin, M.M., 2002. Water Chemistry. Waleland Press Inc., p. 139
Bielski, B.H.J., Cabelli, D.E., Arudi, R.L., Ross, A.B., 1985. Reactivity of HO2/O2 radicals
in aqueous solution. J. Phys. Chem. Ref. Data 14, 1041e1100.
Boltz, D., Holwell, J., 1978. Colorimetric Determination of Nonmetals, second ed.
John Wiley and Sons, New York.
Boonrattanakij, N., Lu, M., Anotai, J., 2009. Kinetics and mechanism of 2,6-dimethy-
aniline degradation by hydroxyl radicals. J. Hazard. Mater. 172, 952e957.
Buxton, G.V., Greenstock, C.L., Helman, W.P., Ross, A.B., 1988. Critical review of rate
constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl
radicals (OH/O) in aqueous solution. J. Phys. Chem. Ref. Data 17, 513e886.
Crittenden, J.C., Hu, S., Hand, D.W., Green, S.A., 1999. A kinetic model for H2O2/UV
process in a completely mixed batch reactor. Wat. Res. 33, 2315e2328.
Davies, M.S., Mile, B., Rowlands, C.C., 1995. Oxidation of 4-methylcatechol by
dioxygen studies by ESR spectroscopy. The different regioselectivity of OH and
MeO nucleophilic attack and kinetic deuterium isotope effects. Magn. Reson.
Chem. 33, 15e19.
De Luis, A., Lombran ~ a, J.I., Menendez, A., 2011. Modeling of the radicalary state in
the H2O2/UV oxidation system to predict the degradation kinetics of phenolic
mixture solutions. AIChE J. 30, 196e207.
Dong, B., Kahl, A., Cheng, L., Vo, H., Ruehl, S., Zhang, T., Snyder, S., S
aez, A.E.,
Quanrud, D., Arnold, R.G., 2015. Fate of trace organics in a wastewater efuent
dependent stream. Sci. Total Environ. 518e519, 479e490.
Eriksen, T.E., Lind, J., Merenyi, G., 1985. On the acid-base equilibrium of the car-
bonate radical. Radiat. Phys. Chem. 26, 197e199.
Ershov, B.G., Janata, E., Alam, M.S., Gordeev, A.V., 2008. Studies of the reaction of the
hydroxyl radical with the oxalate ion in an acidic aqueous solution by pulse
radiolysis. Russ. Chem. Bull. Int. Ed. 57, 1189.
Getoff, N., Schworer, F., Markovie, V.M., Sehested, K., Nielsen, S.O., 1971. Pulse
radiolysis of oxalic acid and oxalates. J. Phys. Chem. 75, 749e755.
Glaze, W.H., Lay, Y., Kang, J.-W., 1995. Advanced oxidation processes. A kinetic
model for the oxidation of 1,2-dibromo-3-chloropropane in water by the
combination of hydrogen peroxide and UV radiation. Ind. Eng. Chem. Res. 34,
2314e2323.
Greenwood, N.N., Earnshaw, A., 1997. Chemistry of the Elements, second ed. But-
terworth-Heinemann, p. 310.
Gohn, M., Getoff, N., 1977. Pulse radiolysis of 3,4-dihydroxytoluene. J. Chem. Soc. 73,
1207e1215.
Hart, E.J., 1951. Mechanism of the g-Ray induced oxidation of formic acid in aqueous
solution. J. Am. Chem. Soc. 73, 68e73.
Holeman, J., Bjergbakke, E., Sehested, K., 1987. The importance of radical-radical
reactions in pulse radiolysis of aqueous carbonate/bicarbonate. Proc. Tihany
Symp. Radiat. Chem. 1, 149e153.
 T. Zhang et al. / Chemosphere 161 (2016) 349e357
Housecroft, C.E., Sharpe, A.G., 2005. Inorganic Chemistry, second ed. Prentice-
Pearson-Hall, p. 368.
Huang, D., Zhang, X., Chen, Z.M., Zhao, Y., Shen, X.L., 2011. The kinetics and
mechanism of an aqueous phase isoprene reaction with hydroxyl radical.
Atoms. Chem. Phys. 11, 7399e7415.
Jin, F., Zhong, H., Cao, J., Cao, J., Kawasaki, K., Kishita, A., Matsumoto, T., Tohji, K.,
Enomoto, H., 2010. Oxidation of unsaturated carboxylic acids under hydro-
thermal conditions. Bioresour. Technol. 101, 7624e7634.
Jongberg, S., Gislason, N.E., Lund, M.N., Skibsted, L.H., Waterhouse, A.L., 2011. Thiol-
quinone adduct formation in myobrillar proteins detected by LC-MS. J. Agric.
Food Chem. 59, 6900e6905.
Joschek, H.I., Miller, S.I., 1966. Photooxidation of phenol, cresols, and dihydrox-
ybenzenes. J. Am. Chem. Soc. 88 (14), 3273e3281.
Kang, N., Lee, D.S., Yoon, J., 2002. Kinetic modeling of Fenton oxidation of phenol
and monochlorophenols. Chemosphere 47, 915e924.
Lay, Y.S., 1989. Oxidation of 1,2-dibromo-3-chloropropane in Groundwater Using
Advanced Oxidation Processes. Ph.D. dissertation. University of California, Los
Angeles.
Lee, P.-C., Lee, W., 1996. In vivo estrogenic action of nonylphenol in immature fe-
male rates. Bull. Environ. Contam. Toxicol. 57, 341e348.
Leitner, N.K.V., Dore, M., 1997. Mechanism of the reaction between hydroxyl radicals
and glycolic, glyoxylic, acetic and oxalic acids in aqueous solution: consequence
on hydrogen peroxide consumption in the H2O2/UV and O3/H2O2 systems.
Water Res. 31, 1383e1397.
Liao, C.-H., Gurol, M., 1995. Chemical oxidation by photolytic decomposition of
hydrogen peroxide. Environ. Sci. Technol. 29, 3007e3014.
Loyo-Rosales, J.E., Rice, C.P., Torrents, A., 2007. Fate of octyl- and nonylphenol
ethoxylates and some carboxylated derivatives in three American wastewater
treatment plants. Environ. Sci. Technol. 41, 6815e6821.
Lye, C.M., Frid, C.L.J., Gill, M.E., Cooper, D.W., Jones, D.M., 1999. Estrogenic alkyl-
phenols in sh tissues, sediments, and waters from the UK. Tyne Tees Estuaries.
33, 1009e1014.
MacDonald, B.C., Lvin, S.J., Patterson, H., 1997. Correction of uorescence inner lter
effects and the partitioning of pyrene to dissolved organic carbon. Anal. Chim.
Acta 338, 155e162.
Matsuura, T., Omura, K., 1974. Photochemical hydroxylation of aromatic com-
pounds. Synthesis 3, 173e184.
Moreira, J., Serrano, B., Ortiz, A., de Lasa, H., 2012. A unied kinetic model for phenol
photocatalytic degradation over TiO2 photocatalysts. Chem. Eng. Sci. 78,
186e203.
Murcia, M.D., Vershinin, N.O., Briantceva, N., Gomez, M., Gomez, E., Cascales, E.,
Hidalgo, A.M., 2015. Development of a kinetic model for the UV/H2O2 photo-
degradation of 2,4-dichlorophenoxiacetic acid. Chem. Eng. J. 266, 356e367.
Nagarnaik, P.M., Boulanger, B., 2011. Advanced oxidation of alkylphenol ethoxylates
in aqueous systems. Chemosphere 85, 854e860.
Neta, P., Huie, R.E., Ross, A.B., 1988. Rate constants for reactions of inorganic radicals
in aqueous solution. J. Phys. Chem. Ref. Data. 17, 1027e1284.
Olariu, R.I., Klotz, B., Barnes, I., Becker, K.H., Mocanu, R., 2002. FT-IR study of the
ring-retaining products from the reaction of OH radicals with phenol, o-, m-,
and p-cresol. Atmos. Environ. 36, 3685e3697.
Omura, K., Matsuura, T., 1968. Photo-induced reactions-IX: the hydroxylation of
phenols by the photo-decomposition of hydrogen peroxide in aqueous media.
357
Tetrahedron 24, 3475e3487.
Perry, R., Green, D., Maloney, J., 1981. Perrys Chemical Engineerings Handbook,
fth ed. McGraw-Hill, New York.
Pospisil, J., Ettel, V., Skola, V., 1957. Pyrocatechol oxidation to muconic acid. Chem.
Prumysl. 7, 244e248.
Rojas, M.R., Perez, F., Whitley, D., Arnold, R.G., S
aea, A.E., 2010. Modeling of
advanced oxidation of trace organic contaminants by hydrogen peroxide
photolysis and Fentons reaction. Ind. Eng. Chem. Res. 49, 11331e11343.
Rojas, M.R., Leung, C., Whitley, D., Zhu, Y., Arnold, R.G., Sa
ea, A.E., 2011. Advanced
oxidation of trace organics in water by hydrogen peroxide solar photolysis. Ind.
Eng. Chem. Res. 50, 12479e12487.
Scheck, C.K., Frimmel, F.H., 1995. Degradation of phenol and salicylic acid by ul-
traviolet radiation/hydrogen peroxide/oxygen. Wat. Res. 29, 2346e2352.
Scullion, S.D., Clench, M.R., Cooke, M., Ashcroft, A.E., 1996. Determination of sur-
factants in surface water by solid-phase extraction, liquid chromatography and
liquid chromatography-mass spectrometry. J. Chromatogr. A 733, 207e216.
Sehested, K., Cortzen, H., Christensen, H.C., Hart, E.J., 1975. Rates of reaction of O,
OH and H with methylated benzenes in aqueous solution. Optical spectra of
radicals. J. Phys. Chem. 79, 310e315.
Soto, A.M., Justicia, H., Wray, J.W., Sonnenschein, C., 1991. p-Nonyl-phenol: an es-
trogenic xenobiotic released from modied polystyrene. Environ. Health
Perspect. 92, 167e173.
Stefan, M.I., Hoy, A.R., Bolton, J.R., 1996. Kinetics and mechanism of the degradation
and mineralization of acetone in dilute aqueous solution sensitized by the UV
photolysis of hydrogen peroxide. Environ. Sci. Technol. 30, 2382e2390.
Song, W., Ravindran, V., Pirbazari, M., 2008. Process optimization using a kinetic
model for the ultraviolet radiation-hydrogen peroxide decomposition of natural
and synthetic organic compounds in groundwater. Chem. Eng. Sci. 63,
3249e3270.
Sun, W., Saeys, M., 2008. First principles study of the reaction of formic and acetic
acids with hydroxyl radicals. J. Phys. Chem. A 112, 6918e6928.
Swartz, C.H., Reddy, S., Benotti, M.J., Yin, H., Barber, L.B., Brownawell, B.J., Rudel, R.A.,
2006. Steroid estrogens, nonylphenol ethoxylate metabolites, and other
wastewater contaminants in groundwater affected by a residential septic sys-
tem on Cape Cod, MA. Environ. Sci. Technol. 40, 4894e4902.
Volman, D.H., Chen, J.C., 1959. The photochemical decomposition of hydrogen
peroxide in aqueous solutions of allyl alcohol at 2537 . J. Am. Chem. Soc. 81
(16), 4141e4144.
Waite, J.H., 1976. Calculating extinction coefcients for enzymatically produced 0-
quinones. Anal. Biochem. 75, 211e218.
Wang, K., Hsieh, Y., Chen, L., 1998. The heterogeneous photocatalytic degradation,
intermediates and mineralization for the aqueous solution of cresols and
nitrophenols. J. Hazard. Mater. 59, 251e260.
Waters, W.A., 1964. Mechanism of Oxidation of Organic Compounds. Methuen,
p. 132.
Weinstein, J., Bielski, B.H.J., 1979. Kinetics of the interaction of perhydroxyl and
superoxide radicals with hydrogen peroxide. The Haber-Weiss reaction. J. Am.
Chem. Soc. 101, 58e62.
Wojna
rovits, L., Foldiak, G., Dangelantonio, M., Emmi, S.S., 2002. Mechanism of OH
radical-induced oxidation of p-cresol to p-methylphenolxyl radical. Res. Chem.
Intermed. 28, 373e386.